Founding Platform & AI Systems Engineer

Most AI systems today were not designed for children. Children are either locked out entirely, or exposed to tools that were never designed for learning environments.

JurneeGo solves this by combining:

• Child-safe AI interaction
• Parent visibility into learning journeys
• Teacher integration with classroom learning
• Compliance with global child data privacy laws

Our mission is to build the technical infrastructure that makes AI safe and usable in education environments.

Tiếng Việt: JurneeGo đang xây dựng nền tảng AI học tập an toàn cho trẻ em, kết nối trẻ em, phụ huynh và giáo viên trong một môi trường AI minh bạch và có kiểm soát. Sứ mệnh của chúng tôi là xây dựng hạ tầng kỹ thuật giúp AI có thể được sử dụng an toàn trong giáo dục.

• Strong system-level thinker, not only feature-focused.
• Experience designing backend or distributed systems.
• Comfortable designing multi-tenant systems and RBAC models.
• Experience working with cloud environments (AWS preferred).
• Deep strength in AI systems OR Platform/DevOps.
• High ownership mindset and comfortable operating with ambiguity.
• AI-native engineer using modern AI coding tools effectively.

Language Requirement / Yêu Cầu Ngôn Ngữ

• English is helpful but not required.
• Technical capability and system thinking are the primary evaluation criteria.

Core role definition

“The System Stabilizer & Scale Architect.” If Vy (FE #1) is speed and product build, then FE #2 is system integrity, scalability, cost control, and safety enforcement.

I. Role mission (non-negotiable)

Primary objective (first 90 days): take what Vy builds and make it reliable, scalable, safe, and cost-efficient — without slowing product velocity.

II. Archetype (who this person is)

You are NOT hiring another frontend builder or another “AI tinkerer.” You ARE hiring a “Builder–Operator Hybrid” — someone who has built systems AND run them in production, thinks in failure modes (not just features), and understands tradeoffs (cost vs latency vs safety).

III. Core competencies

• 1. Backend Systems Ownership (CRITICAL): must be able to own APIs, data flow, event systems (WebSockets / real-time), and session architecture (SSA) — including the Child → AI → Parent → Teacher loop. This is not trivial; it is the core product.
• 2. AI Routing + Cost Optimization (CORE IP): must understand multi-model routing, token cost dynamics, and latency tradeoffs. Example: cheap model → 80% traffic, premium model → 20%, while maintaining UX consistency — A/B model routing and cost-aware scaling. If they don't get this → reject.
• 3. Safety & Compliance Engineering (MANDATORY): they must think “What happens if this goes wrong with a child?” Must understand input/output filtering, guardrails (pre + post LLM), logging + audit trails, and crisis escalation flows — a safety layer BEFORE the model and AFTER the model. This is not optional; this is the moat.
• 4. Real-Time Systems (IMPORTANT): the product requires live parent monitoring, shared sessions, and multi-user sync. They should understand WebSockets, Pub/Sub (Redis), and event-driven architecture.
• 5. Cost Engineering Mindset: they think “every feature = cost per query.” Must be able to model cost per user, optimize infra vs LLM spend, and implement caching strategies.
• 6. Observability & Debugging: they must build logging systems, monitoring dashboards, and failure tracing — because there will be AI unpredictability and safety edge cases.

IV. What they are NOT

Reject if they are: a pure frontend engineer (no); enterprise DevOps only (too slow, too rigid); a research AI engineer (too theoretical); or a “prompt engineer” (useless at this stage).

V. Personality traits (very important)

• Low ego, high ownership: works WITH Vy, not against; doesn't create silos.
• Calm under chaos: early startup is messy; must handle ambiguity.
• Obsession with tradeoffs: not the “best solution” but the “right solution now.”
• Builder mentality: ships, fixes, iterates.

VI. Interaction model (Vy + FE #2)

This is critical. Vy (FE #1) moves fast, builds features, and experiments. FE #2 hardens systems, refactors intelligently, and prevents future collapse. Together: speed + stability.

VII. 30-60-90 day expectations

• First 30 days — Goal: understand the system deeply. Review architecture, map data flow, identify risks. Deliverable: a “System Risk Map.”
• 60 days — Goal: own the infrastructure layer. Improve API structure, optimize AI routing, implement logging + monitoring. Deliverable: a stable backend system.
• 90 days — Goal: become system co-architect. Improve performance, reduce costs, lead backend decisions. Deliverable: a scalable, production-ready system.

VIII. Interview questions (use these)

• System thinking: “Design how a child query flows through JurneeGo.” Look for: safety layer first, routing logic, logging.
• AI routing: “How do you reduce LLM cost by 50% without hurting UX?”
• Failure mode: “What happens if the AI gives a dangerous answer?”
• Real-time systems: “How would you build parent live monitoring?”
• Tradeoffs: “Would you prioritize latency or cost? Why?”

IX. Hiring bar (non-negotiable)

They must: think in systems, understand AI infra, care about safety, be cost-aware, and ship fast. If any are missing → pass.

X. Offer narrative (what you tell them)

You're not hiring an employee. You're recruiting a builder.

Subject: Join JurneeGo as Founding Engineer #2. You won't just write code here. You'll define how a system that children depend on actually works. Every decision you make will affect how safe the product is, how fast it scales, and how much it costs to run. You'll take what we build — and make it real, stable, and world-ready. If that excites you, you'll do the best work of your career here. — Gui

Final insight

Your biggest advantage right now is that you are building BOTH a product AND an architecture moat. FE #1 builds the product. FE #2 protects the moat.